Figure 1.
Misexpression in testes of F1 and F2 hybrids.
(A) Proportions of underexpressed (green) and overexpressed (magenta) transcripts in fertile (DxM) F1s, sterile (MxD) F1s, and F2s. Mean values are shown for F1s and boxplots for F2s indicate median, interquartile range, and outliers >1.5× interquartile range are shown as points. (B) Proportions of misexpressed transcripts common to sterile (MxD) F1s and F2s. (C) Columns indicate proportions of transcripts on each chromosome significantly positively (red, upward) and negatively (blue, downward) correlated with right relative testis weight. (D) QTL mapping of misexpression (number of under- or over-expressed transcripts/individual) in F2s. Significance thresholds, determined by permutation, are indicated with dashed lines. (E) Overexpression of X-linked transcripts in F2s by two-locus genotype for chromosomes 17 (10.4 cM) and X (15.0 cM). Boxes indicate interquartile range, horizontal lines indicate medians, and whiskers extend to 1.5× interquartile range. Outliers are indicated with points.
Table 1.
Expression quantitative trait loci (eQTL).
Figure 2.
Genomic distribution of eQTL and QTT.
Heatmap showing eQTL locations (marker/pseudomarker positions on x-axis) and transcript locations (y-axis). LOD scores above permutation thresholds are shown in dark blue and LOD scores >10 in light blue.
Table 2.
trans eQTL hotspots.
Figure 3.
Chromosome 17 hotspot position and effects implicate Prdm9.
(A) Pie charts showing trans eQTL in the chromosome 17 hotspot are largely under- or over-dominant, in contrast to the pattern seen for trans eQTL overall. (B) Histogram of trans eQTL counts for 4 cM sliding window overlaid with LOD plots for coincident sperm count and testis weight QTL.
Table 3.
Functional annotation of QTT associated with trans eQTL hotspots.
Table 4.
Candidate genes in trans eQTL hotspots.
Table 5.
Conditional mapping results by covariate.
Figure 4.
Genetic interactions revealed by conditional mapping.
Genome plot generated using circos software [106]. Each line represents an interaction eQTL hotspot; color and thickness indicate number of eQTL. Red rectangles indicate sterility QTL positions and dark blue rectangles indicate trans eQTL hotspots (original mapping). Grey triangles indicate positions of marker genotypes used as covariates in conditional mapping.
Figure 5.
Interaction network. Interaction eQTL hotspots identified with different genotype covariates are shown as single nodes if the distance between regions was <12.8 Mb (average distance between genotyping markers).
Nodes are labeled with chromosome, and position (cM) in superscript. Nodes with musculusPWD sterile alleles are magenta, domesticusWSB sterile alleles in blue and sterile heterozygous genotypes in green. Edge weight indicates the number of interaction eQTL. Node size is proportional to total number of interactions. Edge color matches sterile allele at marker used as covariate and arrow points to node of peak position. Edges with two arrowheads indicate reciprocal covariate/peak interactions between nodes; if sterile alleles differ, edge is gray and arrowheads indicate sterile allele at opposite node.
Figure 6.
Model of genome-wide expression effects caused by X-17 interaction.
A musculusPWD allele on the X @15 cM interacts with heterozygous 17@13 cM (likely Prdm9) to cause overexpression of the X chromosome during meiosis. X-linked transcriptional regulation genes, which are usually silenced by MSCI, affect expression of autosomal genes.